Skip to main content

Full text of "Pulsars as gravitational wave detectors"

See other formats


Pulsars as gravitational wave detectors 



George Hobbs 



Abstract Pulsar timing array projects are carrying out high precision observations 
of millisecond pulsars with the aim of detecting ultra-low frequency 10^^ to 
10^^ Hz) gravitational waves. We show how unambiguous detections of such waves 
can be obtained by identifying a signal that is correlated between the timing of 
different pulsars. Here we describe the ongoing observing projects, the expected 
sources of gravitational waves, the processing of the data and the implications of 
current results. 



1 Introduction 

The first strong evidence for the existence of gravitational waves (GWs) was ob- 
tained by measuring the orbital period decay in the PSR B1913+16 binary pulsar 
system ll25l l59l l62ll . Various ground- and space-based detectors have been devel- 
oped with the aim of making a direct detection of GW signals. Unfortunately, to 
date, no such detection has been made. 

Millisecond pulsars are amazingly stable rotators. In many cases a simple model 
for the pulsar spin-down can be used to predict pulse times-of-arrival (TOAs) with 
an accuracy and precision of < 1/is over many years of observation. The pulsar 
timing technique relies on this stability to obtain precise measurements of a pulsar's 
spin, astrometric and orbital parameters. However, in some cases the pulse TOAs 
are not exactly as predicted, suggesting that the model and analysis procedure do 
not parameterise all of the physical effects that are affecting the pulse emission, 
propagation and detection. As the effects of GWs are not included in the analysis, the 
existence of any such waves will induce deviations between the actual and predicted 
TOAs (commonly known as the 'timing residuals'). Here we describe how these 



George Hobbs 

CSIRO Australia Telescope National Facility, PO Box 76 Epping NSW 1710, Australia, e-mail: 
IgeorgeThobbsgcsiro . au| 



1 



2 



George Hobbs 



signals may be unambiguously confirmed as being caused by GWs and provide 
upper bounds on their amplitude. 

In §2 we provide a basic explanation of how pulsar data sets may be used in the 
search for GWs, §3 describes current data sets and highlights some aspects of the 
data that need careful consideration when the data are processed, §4 lists potential 
sources of GWs, §5 describes how data sets may be accessed and §6 highlights 
future possibilities. 



2 Using pulsars to search for GWs 

Pulsar observations lead to measurements of TOAs at an observatory. The TEMP02 
software package ifTOl l23l can be used to convert these TOAs to a time of emis- 
sion by 1) transforming the TOAs to the Solar System bary centre, 2) determining 
excess propagation delays caused by the interstellar medium and, for binary sys- 
tems, 3) transforming to the pulsar frame. The derived time of emission can then be 
compared with a pulsar model to form the timing residuals. 

The induced pulsar timing residuals caused by a GW signal were first calculated 
at the end of the 1970s ||9] |52l. This early work showed that a GW signal causes 
a fluctuation in the observed pulse frequency, 5v/v, which induces pulsar timing 
residuals at time t from the initial observation as 

Rit) = ^f'-^,r. (1) 
Jo V 

The Dopper shift can be shown to have the form (e.g. iflTl ) 

^=H'Jil^j-hf^ (2) 

where /if^ is the GW strain at the Earth at the time of observation, hfj the strain at the 
pulsar when the electromagnetic pulse was emitted and H'^ is a geometrical term 
that depends upon the angle between the Earth, pulsar and GW source. The GW 
strains evaluated at the positions of multiple pulsars will be uncorrected, whereas 
the component at the Earth will lead to a correlated signal in the timing residuals of 
all pulsars. 

Hobbs et al. (2009 ET\ ) demonstrated how pulsar TOAs affected by GWs can 
be simulated using the TEMP02 software package. Since the intrinsic pulsar pulse 
period, spin-down, orbital motion and various astrometric parameters are a priori 
unknown, they need to be determined from the pulsar timing data. Initial estimates 
of the pulsar parameters are used to form pre-fit timing residuals. A least-squares 
fitting procedure is subsequently used to fit an analytical model to obtain improved 
pulsar parameters and "post-fit" timing residuals (see ||23l ). In Figure[T]we plot pre- 
and post-fit timing residuals for a simulated pulsar in the presence of GWs from 1) 



Pulsars as gravitational wave detectors 

J1857 + 0943 (rms = 0.Z54 /js) pre-fit 



J1B57 + 0943 (rms = 0,252 fis) post-fit 



3 




-2000 -1000 1000 2000 

MJD-52003.8 
J1857 + 0943 (rms = 4. 385 fj.s) pre-fit 



-2000 -1000 1000 2000 

MJD-52003.8 
J1B57 + 0943 (rms = 2,786 (is) post-fit 





-1000 -500 500 1000 

MJD-51005,6 
J1857+0943 (rms = 0,712 iis) pre-fit 



-1000 -500 500 1000 

MJD-51005,6 
J1857+0943 (rms = 0,124 /is) post-fit 





-2000 2000 

MJD-52001,B 
J1857 + 0943 (rms = 8,952 iJ.s) pre-fit 



-2000 2000 

MJD-52001,8 
J1B57 + 0943 (rms = 0,662 /2s) post-fit 




-1000 1000 

MJD-52003,8 



-1000 1000 

MJD-52003,8 



Fig. 1 Simulations of pre-fit (left column) and post-fit (right column) timing residuals. The first 
row simulates a non-evolving, black-hole binary system defined by A+ — Ax = 5 x 10^'* and 
0) = 1 X 10^'. The second row contains a simulation of the predicted timing residuals from the 
evolving source 3C66B (see Section|4j. The third row contains a GW burst with memory source 
leading to a change in rotational frequency Av = 1 x 10^'^ Hz and the bottom row is a realisation 
of a GW background defined by a = —2/3 and A = 10^'*. 



4 George Hobbs 




Angular separation (deg) 

Fig. 2 The expected correlation in the timing residuals of pairs of pulsars as a function of angular 
separation for an isotropic GW background (solid line). The points correspond to simulated pulsar 
data sets in the presence of a GW background with power-law index a = +3/2 and amplitude 
A = 0.01; Figure from Hobbs et al. (2009, HTl ). 



a single, non evolving binary black-hole system, 2) an evolving binary black-hole 
system, 3) a burst GW source and 4) an isotropic stochastic background of GWs. 

It is not possible to determine the exact origin of the timing residuals for a single 
pulsar data set. Any observed residuals may have been caused by, for example, 
irregularities in terrestrial time standards |i48|, errors in the planetary ephemeris, 
irregular spin-down of the pulsar 1191 , calibration effects (e.g. 11661 ) or GWs. These 
effects can only be distinguished by searching for correlations in the timing residuals 
of multiple pulsars. For instance, residuals caused by the irregular spin-down of 
one pulsar will be uncorrelated with the residuals observed for a different pulsar. 
Irregularities in terrestrial time standards will lead to correlated residuals for all 
pulsars (this is only true for data sets which have the same data span; see Section^. 
For an isotropic, stochastic GW background, the GW strain at each pulsar will be 
uncorrelated, but the GW strain at the Earth provides a common signal. The common 
signal for such a background was determined by Hellings & Downs (1983) and is 
reproduced in Figure |2] More recently it has been shown that approximately 20 
pulsars are necessary, timed with an rms timing residual of ^ 100 ns or better over 
five years with observations every week, in order to make a significant detection of 
a possible GW background signal Ii28il . 

Pulsar timing experiments are sensitive to GW signals in the ultra-low frequency 
band (f ^ 10^^ to 10^^ Hz) as individual pulsars are typically only observed once 
every few weeks, have maximum data spans of years to decades and require the 




log^o(f/Hz) 

Fig. 3 Characteristic strain sensitivity for existing and proposed GW detectors as a function of 
GW frequency. Predicted signal levels from various astrophysical GW backgrounds are shown. 
The cun'ent limit is that published by Jenet et al. (2006). The 'PPTA limit' is a predicted limit if 
the PPTA project achieves its design sensitivity. The SKA limit is a prediction of the sensitivity of 
the planned Square Kilometre Array telescope. 



fitting of a pulsar timing model. The pulsar timing method is therefore complemen- 
tary to other GW detection methods such as the Laser Interferometer Space Antenna 
(LISAfl and ground based interferometer systems (such as the Laser Interferome- 
ter Gravitational Wave Observatory, LIGCq), which are sensitive to high frequency 
GWs. In Figure|3]we plot the sensitivity of the pulsar timing experiments, LISA and 
LIGO as well as some expected GW sourcesjj Pulsar timing projects that aim to 
obtain data sets on the most stable millisecond pulsars with the aim of searching for 
correlated timing residuals are known as pulsar 'timing arrays'. The first attempts 
to undertake such projects are described by Romani (1989) and Foster & Backer 
(1990). 



' http://lisa.nasa.gov 

^ http://www.ligo.caltech.edu/advLIGO/ 

' Various authors e.g. I35II44I have indicated that it may be possible to limit the existence of GWs 
with even lower frequencies (10^'^ to 10^* Hz) where the minimum sensitive GW frequency is 
determined by the distance to the pulsar. However, the methods required to achieve this sensitivity 
are different to those described in this paper. We also note that a few other methods exist that can 
limit the ultra-low frequency GW background. One example is reported by Gwinn et al. (1997) 
who obtained an upper bound on the background amplitude by analysing limits on proper motions 
of quasars. 



6 



George Hobbs 



3 Current data sets 

The International Pulsar Timing Array (IPTA) project has the main aim of detecting 
GW signals using pulsar observations and is a collaboration between three separate 
projects. The European project (EPTA flSl ) currently obtains data using the Effels- 
berg, Jodrell Bank, Nangay and Westerbork telescopes (a new telescope in Sardinia 
is currently under construction and, when commissioned, will be used as part of 
the EPTA). The North American project (NANOGrav 123) uses the Arecibo and 
Green Bank telescopes and the Parkes project (PPTA; e.g. Il22l [39l |68l ) uses the 
Parkes radio telescope in Australia. In total approximately 37 different pulsars are 
being observed by these project^ (see ifTSl ). Some of the most precise timing resid- 
uals have been obtained for the brightest millisecond pulsar, PSR J0437— 4715. The 
rms timing residuals over ^ 1 yr obtained with the Parkes telescope at an observ- 
ing frequency close to 3 GHz is ^60 ns and some individual TOA uncertainties are 
'^30ns. Achieving such low rms timing residuals relies on removing the effects of 
dispersion measure variations [74], precisely calibrating the data 1671 and determin- 
ing many post-Keplerian binary parameters ||70| . However, over longer data spans 
clear irregularities are observed in the timing residuals (see Figure|4l R. Manchester, 
private communication). These residuals differ from those published by Verbiest et 
al. (2008) as 1) new data have been added between the years 2006 and 2010 and 
2) the time offsets between different backend systems at the observatory have been 
measured and subsequently not included in the fitting procedures. The cause of the 
observed variations is not currently known. 

Pulsars exhibit two main types of timing irregularity; 'glitches' in which the 
pulsar's rotation rate suddenly increases before undergoing a period of relaxation 
and 'timing noise' which consists of low-frequency features in the timing residu- 
als. Hobbs et al. (2010) analysed a sample 366 pulsars and showed that the tim- 
ing noise, in general, consisted of quasi-periodic oscillations with periodicities of 
years to decades. However, the "amount" of timing noise decreases with charac- 
teristic age and millisecond pulsars, which are used in timing array experiments, 
are typically extremely stable. In a few cases, timing noise and gUtches are clearly 
seen in the residuals for millisecond pulsars. For instance, the timing residuals for 
PSR J1939+2134 have long been known (e.g. l33\) to be dominated by timing 
noise. The first glitch in a millisecond pulsar was found in PSR B 182 1—24 which 
is located in the globular cluster M28 161. The glitch was shown to follow the main 
characteristics of glitches seen in the slower pulsars. Verbiest et al. (2009) studied 
the timing stability for 20 millisecond pulsars included in the PPTA project, over 
time scales up to 10 yr Apart from PSR J1939+2134 it was concluded that the tim- 
ing of most of the other pulsars is stable enough for GW detection on decadal time 
scales. 



* The exact number of pulsars varies as new pulsars are discovered and added to the IPTA sam- 
ple. Radio pulsar surveys are being earned out (Keith et al., submitted to MNRAS,17|) and the 
Fermi gamma-ray spacecraft, has recently discovered numerous previously unidentified gamma- 
ray sources. Many of these have now been shown to be millisecond pulsars (R Ray, these proceed- 
ings) some of which are likely soon to be included in the IPTA. 



Pulsars as gravitational wave detectors 

0437-4715 (rms = 0.433 fis) pre-fit 



7 




1995 2000 2005 2010 



Year 

Fig. 4 The timing residuals for PSR J0437— 4715 over 14 years. 



The basic method to detect a GW background signal in pulsar data is to search 
for correlated timing residuals between different pairs of pulsars. However, we note 
that these correlations only hold in the case where all pulsar data sets are processed 
using the same fitting procedure and have the same data span. In Figure |5] post-fit 
timing residuals are shown that include simulated clock irregularities. Even though 
the same signal has been added to each data set, the residuals are not highly corre- 
lated because of the differing data spans and the model fits that have been carried 
out. The following need to be taken into account when attempting to detect or limit 
GW signals in pulsar timing data: 

• Pulsar timing residuals are irregularly sampled, have large gaps and different 
pulsars have different sampling and data spans. 

• The uncertainty on each TOA for a given pulsar can vary by an order-of- 
magnitude because of interstellar scintillation, different observing durations for 
different observations and when combining data sets obtained from different ob- 
servatories. For an unknown reason it is often found that the TOA uncertainties 
are under- or over-estimated (by factors of ^2). 

• Timing residuals are always related to a pulsar timing model for which various 
fits have been carried out to determine the pulsar's spin, astrometric and orbital 
parameters. 

• Arbitrary offsets can exist between TOAs measured with different instruments or 
at different observatories. 

• The timing residuals may be affected by timing noise or glitch events. 



8 



George Hobbs 



psrl (rms = 0.274 ^s) post-fit psr2 (rms = 0,170 /^s) post— fit 




1990 1995 2000 2005 2010 



Year 

Fig. 5 Post-fit timing residuals for three simulated pulsars. Each pulsar data set consists of white 
noise plus a clock irregularity that is the same for each pulsar. The post-fit residuals are not corre- 
lated because of model fitting and because each data set has a different data span. 

4 Potential sources of gravitational waves 

Theoretical models predict that pulsar timing experiments are sensitive to individual, 
periodic sources of GWs, burst sources and a background of GWs. This section 
describes the various possible sources that have been discussed in the literature. 



4.1 Single sources 

GWs are generated by the acceleration of massive objects. For a two-body orbital 
system the GW frequency is twice the orbital frequency implying that supermas- 
sive binary black-hole systems can produce GW signals that are strong enough, and 
are in the correct frequency regime, to be detectable by pulsar timing experiments. 
An estimate of the timing residual for a massive black-hole binary system can be 
obtained from ||271 : 



Pulsars as gravitational wave detectors 



9 



where d is the distance to the black-hole binary which has a total mass of M/(l + z) 
and emits GWs at frequency /. 

Sesana, Vecchio & Volonteri (2009) showed that it is likely that the maximum 
timing residuals induced by such a system will be in the '--^ 5 —50 ns range. Any 
such source is likely to be from a massive system (with a chirp mass Mc > 5 x 
10^ M0) and at a redshift between 0.2 < z < 1 .5. More recently Sesana & Vecchio 
(2010) showed for a "fiducial" timing array project containing 100 pulsars uniformly 
distributed in the sky that, if a source is detected with a signal to noise ratio of 10, the 
source position could be determined to within ^ 40 deg^ with a fractional error on 
the signal amplitude around 30% and the source inclination and polarisation angles 
could be recovered to within ~ 0.3 rad. 

Unfortunately there has not yet been any clear detection of a supermassive binary 
black hole system where the black holes are close enough, and massive enough, for 
the system to be emitting detectable GW emission. Rodriguez et al. (2006) dis- 
covered a system in the radio galaxy 0402+379, which has a projected separation 
between the two black holes of just 7.3 pc. However, this is still too wide to be 
emitting detectable GWs. Sudou et al. (2003) reported the possible detection of a 
supermassive black hole binary in the radio galaxy 3C 66B and provided the or- 
bital parameters for the postulated system. Jenet et al. (2004) showed that such a 
system should have produced detectable GW emission and managed to rule out its 
existence with high confidence. Lommen & Backer (2001) unsuccessfully searched 
for GW emission from Sagittarius A* which had also been suggested to be part of 
a binary system. Sillanpaa et al. (1996) have identified a candidate supermassive 
black hole binary system in the blazar OJ287 with member masses of 1 .3 x 10^ M0 
and 1.8 X IO'^Mq. However, Yardley et al. (2010) showed that this system will 
require the sensitivity of future telescopes (such as the Square Kilometre Array, 
SKA) for detection. Searches for more supermassive binary black hole candidates 
are ongoing. Wen, Liu & Han (2009) selected 1209 pairs of galaxies from the Sloan 
Digital Sky Survey and searched for features indicative of merging events. More 
recently Burke-Spolaor (submitted to MNRAS) used archival VLBI data to search 
3114 radio-luminous active galactic nuclear for binary supermassive black holes. 
Only one such source (the same as that previously discovered by Rodriguez et al. 
2006) was detected. 

Wen et al. (submitted to ApJ) described the basic methods in which existing 
pulsar timing observations can be used to place constraints on the coalescence rate 
of binary supermassive black holes in the Universe. Yardley et al. (2010) determined 
the sensitivity of the Verbiest et al. (2008, 2009) data sets to individual GW sources 
and obtained a sky-averaged constraint on the merger rate of nearby (z < 0.6) black- 
hole binaries in the early phases of coalescence with a chirp mass of 10'" M,., of 
less than one merger every seven years. This work is summarised in Figure |6] which 
plots the current sensitivity to single sources and predicted future sensitivity (with 
the Square Kilometre Array, SKA, telescope). 



10 



George Hobbs 



o 




o 




Coalescing binary 
black— holes 



o 

C\2 - 



LO 



PTA LISA 



LIGO 



1-10 



8 



6 



4 



2 







2 



4 



logi„(f/Hz) 



Fig. 6 Figure from Yardley et al. (2010) showing the sensitivity of current and future GW obser- 
vatories to individual, periodic GW sources as a function of the GW frequency. 

4.2 Burst sources 

Sources of detectable burst GW emission include 1) the formation of supermassive 
black holes which leads to a day-long burst of radiation 1631 . 2) highly eccentric su- 
permassive black hole binaries lfT2l . 3) close encounters of massive objects 1*341 and 
4) cosmic string cusps ISj. Pshirkov & Tuntsov (2010) constrained the cosmologi- 
cal density of cosmic string loops using photometry and from pulsar timing showing 
that the pulsar timing data provides the most stringent constraints on the abundance 
of light strings. 

Seto (2009), Pollney & Reisswig (2010), van Haasteren & Levin (2010) and 
Pshirkov, Baskaran & Postnov (2010) all considered GW bursts "with memory". 
In general such events consist of a rise in the GW field, followed by an oscillatory 
behaviour for a few cycles and finally convergence to a non-zero value. The per- 
manent change is known as the burst's "memory". Such events can occur during 
close encounters of massive bodies on hyperbolic trajectories or in an asymmetric 
supernova. Pshirkov, Baskaran & Postnov (2010) showed that one such event could 
be detected from distances of up to 1 Gpc (for the case of equal mass supermas- 
sive binary black hole systems of M = 10^ M^^,). A similar event can occur if the 
line of sight to a pulsar passes a cosmic string (Pshirkov & Tuntsov 2010). Such an 
event would instantaneously change the apparent frequency, v, of the pulsar by a 
small amount 4 v in the same way as that of a GW burst with memory although the 



Pulsars as gravitational wave detectors 



11 




a 

Fig. 7 Upper bounds from Jenet et al. (2006) for stochastic backgrounds with various values of the 
GW spectral exponent a (solid line). The dotted lines indicate regions predicted by various GW 
background models. The dashed line indicates potential future limits from the PPTA project. The 
star symbol provides the upper bound given by Kaspi, Taylor & Ryba (1994). Figure from Jenet et 
al. (2006). 



memory phenomenon would apply to all observed pulsars whereas this line-of-sight 
effect is specific to an individual pulsar. 

Amaro-Seoane et al. (2010) discussed the possibihty that supermassive black 
holes do not coalesce before the merger with a third galaxy. This leads to orbits with 
high eccentricity leading to an intense burst of GWs. They showed that, for rea- 
sonable supermassive black-hole evolution models, a few bursts will induce timing 
residuals > 1 ns, however, if pulsar timing were sensitive at the nanosecond level, 
most of these bursts would be undetectable because of confusion with the GW back- 
ground. 

A detection of a burst source using pulsar timing would provide the ability to lo- 
calise the position of the source with the hope of identifying an electromagnetic 
counterpart. The angular resolution of a given timing array experiment depends 
upon the number of pulsars in the array, the timing precision and the distribution 
of the pulsars on the sky. 



4.3 Stochastic background 

A stochastic background of GWs is expected from the early phases of coalescence 
for supermassive black holes ll47ll26ll72l[lTll53l . cosmic strings or relic GWs from 
the big bang 1381 . In most models, the GW strain spectrum, lic{f), is represented by 
a power-law in the GW frequency, /, 



12 



George Hobbs 



hAf)=A^j^y (4) 

where the spectral exponent, a = —2/3, —1 and —7/6 for likely GW backgrounds 
caused by coalescing black hole binaries, cosmic strings and relic GWs respectively. 
The energy density of the background per unit logarithmic frequency interval can 
be written as 

^GM = l^fhc{f? (5) 

where Ho is the Hubble constant. Sesana, Vecchio & Colacino (2008) showed that 
the expected form of the stochastic background is not simply he °= f^'^^^, but the 
frequency dependence becomes steeper above ^ 10^*^ Hz. They showed that the 
major contributors to the background come from massive > IO^^Mq binary black 
hole systems that are relatively nearby (z < 2). 

In contrast to backgrounds formed from supermassive black-hole binary sys- 
tems, Saito & Yokoyama (2009) showed that the formation of intermediate-mass 
(^600 Mg) black holes at the end of the inflationary era will lead to a potentially 
detectable background of GWs. Current pulsar upper bounds already limit the pro- 
cess for the formation of such black holes. 

Both "kinks" and "cusps" on cosmic string loops can also contribute to a back- 
ground of GWs. Olmez, Mandic & Siemens (2010) showed that both phenomena 
contribute at the same order to the background. The most recent constraints, and 
their implications, on the cosmic string tension using pulsar timing experiments 
have been provided by Battye & Moss (2010). 

A background of GW radiation from the early universe is also expected, if de- 
tected, to provide a unique view of the physics of the very early universe (e.g. ifTsl ). 
Boyle & Buonanno (2007) showed that combinations of cosmic microwave back- 
ground experiments and GW detection experiments lead to strong constraints on the 
existence and properties of various energy components that may have dominated the 
universe at the end of the inflationary era. 

One of the first limits on a GW background was placed by Stinebring et al. 
(1990). Kaspi, Taylor & Ryba (1994) used observations of PSRs J1857H-0943 and 
J1939H-2134 from the Arecibo observatory to provide an upper bound on the GW 
background. Their data sets were made publically available allowing other algo- 
rithms to be applied to their data sets 1401 |29l . Jenet et al. (2006) developed a method 
that took into account all the effects of fitting a pulsar timing model to data sets of 
different data spans and applied the technique to carefully selected PPTA data sets 
as the technique can only be applied to data sets that are statistically "white". Van 
Haasteren et al. (2008) developed a Bayesian algorithm which has been applied to 
EPTA data sets [lT3i . Anholm et al. (2009) developed an independent method which 
currently does not address some issues relating to real pulsar data. Work is ongoing 
to develop a method that correctly detects the GW background or limits its ampli- 
tude and can easily be applied to the IPTA data sets. 

The most stringent limits published to date on the amplitude of a background 
caused by these various phenomena were obtained using observations from the 



Pulsars as gravitational wave detectors 



13 



PPTA project combined with archival data from the Arecibo telescope ||29l . These 
upper bounds on A are shown in Figure Q These results were shown to 1) constrain 
the merger rate of supermassive black hole binary systems at high redshift, 2) rule 
out some relationships between the black hole mass and the galactic halo mass, 3) 
constrain the rate of expansion in the inflationary era and 4) provide an upper bound 
on the dimensionless tension of a cosmic string background. 



5 Accessing pulsar data sets 

The IPTA pulsar data sets contain some of the most precisely measured observations 
yet made of radio pulsars. These data sets can be used to search for GW sources and 
for numerous other purposes. Details of the only currently publically available data 
sets are provided in Kaspi, Taylor & Ryba (1994). Hobbs et al. (2009) provided a 
sample of simulated data sets for comparison of GW detection algorithms that are 
available for download. In the near future it is expected that all observations from 
Parkes Observatory (after an embargo period of 18 months) will be available for 
download providing access to most PPTA data. A few data sets relating to the PPTA 
project are also available as part of the PULSE@Parkes outreach proiect ll20ll24l . 

The TEMP02 package provides a few "plugin" packages and routines suitable for 
GW analysis: 

• GWsiM.H: a library that implements the mathematics described by Hobbs et al. 
(2009). The routines in this library can be used to create new TEMP02 plugins 
easily. 

• FAKE: plugin that simulates pulsar TOAs given a timing model. 

• GWsiNGLE: plugin that simulates a single non-evolving GW source and pro- 
duces a file of site-arrival-times that can be used for subsequent processing. 

• GWevolve: plugin that simulates a single evolving GW source. 

• GWbkgrd: plugin that simulates a background of GW sources. 

A new plugin is currently under development that will either provide limits on the 
amplitude of the GW background or give the significance of any detection by search- 
ing for correlations in the timing residuals of multiple pulsars. 



6 The future 

It is likely that the signature of a GW background will be discovered in IPTA data 
sets shortly. However, any such initial "detection" will only have a low significance 
(e.g. (7 ~ 2 to 3). How this significance will increase with time will depend upon 
the intrinsic pulsar timing noise, the amplitude and spectral exponent of the back- 
ground, whether new pulsars are discovered and whether observing systems are 
improved. A very high significance detection, enabling detailed studies of the astro- 



14 



George Hobbs 



physical objects that form the background and the physics of the GWs themselves 
(e.g. ||36l ). will require large numbers of pulsars to be observed with high precision. 
Various new telescopes are being developed that should detect a large number of 
new pulsars or be able to observe pulsars with high precision (e.g. the Australian 
Square Kilometre Array Pathfinder Ellllll, MeerKAT E) and FAST Ell). These 
telescopes are pathfinders to the Square Kilometre Array which should be fully op- 
erational by 2024. This telescope should revolutionise pulsar and GW astronomy by 
allowing almost all pulsars in our Galaxy to be detected and GW detection using 
pulsar observations should become commonplace. 



7 Conclusion 

Pulsar timing experiments can detect gravitational wave signals. The most likely 
detectable source is an isotropic, stochastic background from a large number of 
coalescing supermassive binary black hole systems. It is possible that a detection 
of such a background could be made with existing data, but it will require a further 
5-lOyr to have a definitive, highly significant detection. It is expected that GW 
astronomy using pulsars will become a standard astronomical tool during the Square 
Kilometre Array era when the background will be analysed in detail and individual 
binary black-hole systems and burst GW sources will be detectable. 

Acknowledgements This work is undertaken as part of the Parkes Pulsar Timing Array project, 
which is a collaboration between CSIRO Astronomy and Space Science, Swinburne University, 
the University of Texas, Brownsville and the University of California, San Diego. The Parkes radio 
telescope is part of the Australia Telescope, which is funded by the Commonwealth of Australia 
for operation as a National Facility managed by CSIRO. GH thanks D. Yardley, J. Verbiest, R. 
Manchester and A. Sesana for comments on early drafts of this text. 



References 

1. Amaro-Seoane, P., et al.: Triplets of supermassive black holes: astrophysics, gravitational 
waves and detection. MNRAS 402, 2308-2320 (2010). 

2. Anholm, M., Ballmer, S., Creighton, J.D.E., Price, L.R., Siemens, X.: Optimal strategies for 
gravitational wave stochastic background searches in pulsar timing data. Phys. Rev. D 79(8), 
084, 030 (2009). 

3. Battye, R., Moss, A.: Updated constraints on the cosmic string tension. ArXiv e-prints 
1005.0479 (2010) 

4. Booth, R.S., de Blok, W.J.G., Jonas, J.L., Fanaroff, B.: MeerKAT Key Project Science, Spec- 
ifications, and Proposals. ArXiv e-prints 0910.2935 (2009) 

5. Boyle, L.A., Buonanno, A.: Relating gravitational wave constraints from primordial nucle- 
osynthesis, pulsar timing, laser interferometers, and the CMB: impUcations for the early uni- 
verse. ArXiv e-prints 708 (2007) 

6. Cognard, I., Backer, D.C.: A microglitch in the millisecond pulsar psr bl821-24 in m28. ApJ 
612, L125-L127 (2004) 



Pulsars as gravitational wave detectors 



15 



7. Cordes, J.M., et al.: Arecibo Pulsar Survey Using ALFA. 1. Survey Strategy and First Discov- 
eries. ApJ 637, 446^55 (2006). 

8. Damour, T., Vilenkin, A.: Gravitational wave bursts from cusps and kinks on cosmic strings. 
Phys. Rev. D 64(6), 064,008 (2001) 

9. Detweiler, S.: Pulsar timing measurements and the search for gravitational waves. ApJ 234, 
1100(1979) 

10. Edwards, R.T., Hobbs, G.B., Manchester, R.N.: TEMP02, a new pulsar timing package - II. 
The timing model and precision estimates. MNRAS 372, 1549-1574 (2006). 

11. Enoki, M., Inoue, K.T., Nagashima, M., Sugiyama, N.: Gravitational Waves from Supermas- 
sive Black Hole Coalescence in a Hierarchical Galaxy Formation Model. ApJ 615, 19-28 
(2004) 

12. Enoki, M., Nagashima, M.: The Effect of Orbital Eccentricity on Gravitational Wave Back- 
ground Radiation from Supermassive Black Hole Binaries. Progress of Theoretical Physics 
117, 241-256 (2007) 

13. Ferdman, R.D., et al.: The european pulsar timing array: current efforts and a leap toward the 
future. Classical and Quantum Gravity 27(8), 084,014 (2010). 

14. Foster, R.S., Backer, D.C.: Constructing a pulsar timing array. ApJ 361, 300 (1990) 

15. Grishchuk, L.P: Relic gravitational waves and cosmology. Phys. Uspekhi pp. 1235-1247 
(2005) 

16. Gwinn, C.R., Eubanks, T.M., Pyne, T., Birkinshaw, M., Matsakis, D.N.: Quasar Proper Mo- 
tions and Low-Frequency Gravitational Waves. ApJ 485, 87 (1997). 

17. Hellings, R.W., Downs, G.S.: Upper limits on the isotropic gravitational radiation background 
from pulsar timing analysis. ApJ 265, L39 (1983) 

18. Hobbs, G., et al.: The international pulsar timing array project: using pulsars as a gravitational 
wave detector. Classical and Quantum Gravity 27(8), 084,013 (2010). 

19. Hobbs, G., Lyne, A.G., Kramer, M.: An analysis of the timing irregularities for 366 pulsars. 
MNRAS 402, 1027-1048 (2010). 

20. Hobbs, G., et al.: The PULSE@Parkes Project: a New Observing Technique for Long-Tenn 
Pulsar Monitoring. PASA 26, 468-475 (2009). 

21. Hobbs, G., et al: TEMP02: a new pulsar timing package - 111. Gravitational wave simulation. 
MNRAS 394, 1945-1955 (2009). 

22. Hobbs, G.B., et al.: Gravitational-Wave Detection Using Pulsars: Status of the Parkes Pulsai' 
Timing Array Project. PASA 26, 103-109 (2009). 

23. Hobbs, G.B., Edwards, R.T., Manchester, R.N.: Tempo2, a new pulsar-timing package - i. an 
overview. MNRAS 369, 655-672 (2006). 

24. Hollow, R., et al.: PULSE@Parkes: Pulsar Observing for High School Students. In: 
M. G. Gibbs, J. Barnes, J. G. Manning, & B. Partridge (ed.) Astronomical Society of the 
Pacific Conference Series, Astronomical Society of the Pacific Conference Series, vol. 400, 
pp. 190 (2008) 

25. Hulse, R.A., Taylor, J.H.: Discovery of a pulsar in a binary system. ApJ 195, L51-L53 (1975) 

26. Jaffe, A.H., Backer, D.C.: Gravitational waves probe the coalescence rate of massive black 
hole binaries. ApJ 583, 616-631 (2003) 

27. Jenet, F, et al.: The North American Nanohertz Observatory for Gravitational Waves. ArXiv 
e-prints 0909.1058 (2009) 

28. Jenet, F.A., Hobbs, G.B., Lee, K.J., Manchester, R.N.: Detecting the Stochastic Gravitational 
Wave Background Using Pulsar Timing. ApJ 625, L123-L126 (2005) 

29. Jenet, F.A., et al.: Upper Bounds on the Low-Frequency Stochastic Gravitational Wave Back- 
ground from Pulsar Timing Observations: CuiTent Limits and Future Prospects. ApJ 653, 
1571-1576 (2006). 

30. Jenet, F.A., Lommen, A., Larson, S.L., Wen, L.: Constraining the properties of supermassive 
black hole systems using pulsar timing: Application to 3c 66b. ApJ 606, 799-803 (2004) 

31. Johnston, S., et al.: Science with the Australian Square Kilometre Array Pathfinder. PASA 24, 
174-188 (2007) 

32. Johnston, S., et al. Science with ASKAP. The Australian square-Jdlometre-aiTay pathfinder. 
Experimental Astronomy 22, 151-273 (2008). 



16 



George Hobbs 



33. Kaspi, V.M., Taylor, J.H., Ryba, M.: High-precision timing of millisecond pulsars. III. Long- 
term monitoring of PSRs B1855-F09 and B1937-F21. ApJ 428, 713-728 (1994) 

34. Kocsis, B., Caspar, M.E., Marka, S.: Detection rate estimates of gravity waves emitted during 
parabolic encounters of stellar black holes in globular clusters. ApJ 648, 41 1-429 (2006). 

35. Kopeikin, S.M.: Binary Pulsars as Detectors of Ultra-Low Frequency Graviational Waves. 
Phys. Rev. D 56, 4455 (1997) 

36. Lee, K.J., Jenet, F.A., Price, R.H.: Pulsar Timing as a Probe of Non-Einsteinian Polarizations 
of Gravitational Waves. ApJ 685, 1304-1319 (2008). 

37. Lommen, A.N., Backer, D.C.: Using pulsars to detect massive black hole binaries via gravita- 
tional radiation: Sagittarius A* and nearby galaxies. ApJ 562, 297-302 (2001) 

38. Maggiore, M.: Gravitational wave experiments and early universe cosmology. Phys. Rep. 331, 
283-367 (2000) 

39. Manchester, R.N.: Detection of Gravitational Waves using Pulsar Timing. ArXiv e-prints 
1004.3602 (2010) 

40. McHugh, M.P., Zalamansky, G., Vemotte, F., Lantz, E.: Pulsar timing and the upper limits on 
a gravitational wave background: A Bayesian approach. Phys. Rev. D 54, 5993-6000 (1996) 

41. Nan, R.D., Wang, Q.M., Zhu, L.C., Zhu, W.B., Jin, C.J., Gan, H.Q.: Pulsar Observations with 
Radio Telescope FAST. Chin. J. Atron. Astrophys., Suppl. 2 6, 304-310 (2006) 

42. Olmez, S., Mandic, V., Siemens, X.: Gravitational-Wave Stochastic Background from Kinks 
and Cusps on Cosmic Strings. ArXiv e-prints 1004.0890 (2010) 

43. PoUney, D., Reisswig, C: Gravitational memory in binary black hole mergers. ArXiv e-prints 
1004.4209 (2010) 

44. Pshirkov, M.S.: Investigating ultra-long gravitational waves with measurements of pulsar ro- 
tational parameters. MNRAS 398, 1932-1935 (2009). 

45. Pshirkov, M.S., Baskaran, D., Postnov, K.A.: Observing gravitational wave bursts in pulsar 
timing measurements. MNRAS 402, 417^23 (2010). 

46. Pshirkov, M.S., Tuntsov, A.V.: Local constraints on cosmic string loops from photometry and 
pulsar timing. Phys. Rev. D 81(8), 083,519 (2010). 

47. Rajagopal, M., Romani, R.W.: Ultra-Low-Frequency Gravitational Radiation from Massive 
Black Hole Binaries. ApJ 446, 543-549 (1995) 

48. Rodin, A.E.: Optimal filters for the construction of the ensemble pulsar time. MNRAS 387, 
1583-1588 (2008). 

49. Rodriguez, C, Taylor, G.B., Zavala, R.T, Peck, A.B., Pollack, L.K., Romani, R.W.: A Com- 
pact Supermassive Binary Black Hole System. ApJ 646, 49-60 (2006). 

50. Romani, R.W.: Timing a millisecond pulsar array. In: H. Ogelman, E.P.J, van den Heuvel 
(eds.) Timing Neutron Stars, pp. 113-1 17 (1989) 

51. Saito, R., Yokoyama, J.: Gravitational-Wave Background as a Probe of the Primordial Black- 
Hole Abundance. Phys. Rev. Lett. 102(16), 161,101 (2009). 

52. Sazhin, M.V.: Sov. Astron. 22, 36 (1978) 

53. Sesana, A., Haardt, F, Madau, P., Volonteri, M.: Low-Frequency Gravitational Radiation from 
Coalescing Massive Black Hole Binaries in Hierarchical Cosmologies. ApJ 611, 623-632 
(2004). 

54. Sesana, A., Vecchio, A.: Measuring the parameters of massive black hole binary systems with 
pulsar timing an'ay observations of gravitational waves, prd 81(10), 104,008 (2010). 

55. Sesana, A., Vecchio, A., Colacino, C.N.: The stochastic gravitational-wave background from 
massive black hole binary systems: implications for observations with Pulsar Timing Arrays. 
MNRAS 390, 192-209 (2008). 

56. Sesana, A., Vecchio, A., Volonteri, M.: Gravitational waves from resolvable massive black 
hole binary systems and observations with Pulsar Timing Arrays. MNRAS 394, 2255-2265 
(2009). 

57. Seto, N.: Search for memory and inspiral gravitational waves from supermassive binary black 
holes with pulsar timing arrays. MNRAS 400, L38-L42 (2009). 

58. Sillanpaa, A., et al.: Confirmation of the 12-year optical outburst cycle in blazar OJ 287. A&A 
305, L17 (1996) 



Pulsars as gravitational wave detectors 



17 



59. Smarr, L.L., Blandford, R.: The binary pulsar: Physical processes, possible companions and 
evolutionary histories. ApJ 207, 574-588 (1976) 

60. Stinebring, D.R., Ryba, M.F., Taylor, J.H., Romani, R.W.: Cosmic gravitational-wave back- 
ground: Limits from millisecond pulsar timing. Phys. Rev. Lett. 65, 285-288 (1990) 

61. Sudou, H., Iguchi, S., Murata, Y., Taniguchi, Y.: Orbital Motion in the Radio Galaxy 3C 66B: 
Evidence for a Supermassive Black Hole Binary. Science 300, 1263-1265 (2003). 

62. Taylor, J.H., Weisberg, J.M.: A new test of general relativity: Gravitational radiation and the 
binaiy pulsar PSR 1913+16. ApJ 253, 908-920 (1982) 

63. Thome, K.S., Braginskii, V.B.: Gravitational-wave bursts from the nuclei of distant galaxies 
and quasars - proposal for detection using doppler tracking of interplanetary spacecraft. ApJ 
204, L1-L6 (1976) 

64. van Haasteren, R., Levin, Y.: Gravitational-wave memory and pulsai' timing arrays. MNRAS 
401, 2372-2378 (2010). 

65. van Haasteren, R., Levin, Y, McDonald, P., Lu, T.: On measuring the gravitational- wave back- 
ground using pulsar timing arrays, mnras 395, 1005-1014 (2009). 

66. van Straten, W. : Radio astronomical polarimetry and high-precision pulsar timing. ApJ 642, 
1004-1011 (2006). 

67. van Straten, W., Manchester, R.N., Johnston, S., Reynolds, J.: PSRCHIVE and PSRFITS: 
Definition of the Stokes Parameters and Instrumental Basis Conventions. PASA, 27, 104 
(2009). In press 

68. Verbiest, J.P.W., et al.: Status update of the parkes pulsar timing array. Classical and Quantum 
Gravity 27(8), 084,015 (2010). 

69. Verbiest, J.P.W. et al.: Timing stability of millisecond pulsars and prospects for gravitational- 
wave detection. MNRAS 400, 951-968 (2009). 

70. Verbiest, J.P.W., et al.: Precision timing of PSR J0437-4715: an accurate pulsar distance, a 
high pulsar mass and a limit on the variation of Newton's gravitational constant. ApJ 679, 
675-680 (2008). 

71. Wen, Z.L., Liu, F.S., Han, J.L.: Mergers of luminous early-type galaxies in the local universe 
and gravitational wave background. ApJ 692, 51 1-521 (2009). 

72. Wyithe, J.S.B., Loeb, A.: Low-Frequency Gravitational Waves from Massive Black Hole Bi- 
naries: Predictions for LISA and Pulsar Timing Arrays. ApJ 590, 691-706 (2003) 

73. Yardley, D.R.B., et al.: The Sensitivity of the Parkes Pulsar Timing Array to Individual Sources 
of Gravitational Waves. ArXiv e-prints 1005.1667 (2010) 

74. You, X.P., et al.: Dispersion measure variations and their effect on precision pulsar timing. 
MNRAS 378, 493-506 (2007)